In an earlier post, I addressed the question whether the Prime Minister must be a member of the House of Commons. Here, I reflect briefly on some interesting points that have been raised in response to my original piece, which raise fundamental questions about the nature of constitutional conventions, including the appropriate extent of reliance on historical precedent when attempting to discern the content of the contemporary constitution.
Prompted by Andy Burnham’s attempt to return to Parliament by way of the Makerfield by-election, I argued in a recent post that, notwithstanding isolated 20th century precedents to the contrary, we can now say with some confidence that constitutional convention requires the premier to be a member of the House of Commons. In a thoughtful and carefully argued piece, Dane Luo advances a contrary view, contending that while it is “imperative that the Prime Minister take steps to become a MP”, the “historical practice does seem to tolerate a person being appointed as Prime Minister before winning a general election or by-election”. This raises interesting questions both about the specific content of the current convention governing prime ministerial appointments and, more generally, about the role – and limits – of historical precedent in shaping contemporary understandings of constitutional conventions.
Luo’s argument invokes historical practice in the UK along with historical and current practice elsewhere. For instance, he points to the fact that Mark Carney became Prime Minister of Canada approximately six weeks before winning a seat in the House of Commons. Luo also observes that Patrick Gordon Walker was appointed the UK’s Foreign Secretary in October 1964 even though he had lost his parliamentary seat at the general election – a role that Gordon Walker held until he resigned upon losing a by-election in January 1965. Without doubt, these examples establish that it is possible in Canada today for someone to be appointed Prime Minister in anticipation of their soon being elected an MP, and that it was possible in the UK more than half a century ago for someone to be appointed as a Cabinet Minister in the expectation that they would soon secure a seat in the House of Commons via a by-election triggered for that purpose. But do they establish that it would be consistent with constitutional convention in the UK today for someone to be appointed Prime Minister who was not already a member of the House of Commons?
There are, of course, some grounds for arguing that the answer to that question is “yes”. After all, sixty years is not such a long time in the long history of the UK constitution. And, as I noted in my earlier post, there is an even more germane example of similar vintage: Alec Douglas-Home, a member of the House of Lords, took over as Prime Minister in 1963 following the sudden resignation, through ill-health, of Harold Macmillan, only becoming an MP three weeks later, having resigned his peerage and won a by-election. Meanwhile, it might be thought that the degree of similarity between Canada’s constitutional arrangements and the UK’s is sufficient to render unduly technical any claim that Carney’s recent appointment ahead of his becoming an MP cannot serve as a precedent in the British context.
From a different perspective, however, it can be argued that what is actually unduly technical, or legalistic, is an approach that places so much reliance on precedent that other crucial issues pertaining to the development and identification of constitutional conventions become obscured. In articulating his tripartite test for conventions – concerning the existence of precedent, a sense of obligation on the part of constitutional actors to abide by the precedent, and an underlying normative justification for the relevant practice – Sir Ivor Jennings doubtless supplied a valuable framework. But it is a mistake to suppose that the three elements of his tripartite test are made equal. Indeed, precedent is arguably the least important of the three. That much is apparent from the ease with which it has been possible for some conventions (or apparent conventions) to be departed from with little or no objection. (Take, for instance, Tony Blair’s decision to scrap nearly 40 years of precedent, whereby the Prime Minister had answered oral questions twice a week in the Commons, replacing it with the now well-established once-a-week practice.) Moreover, some conventions – the Sewel convention, for example, concerning Westminster legislation on devolved matters – have emerged with considerable speed, again indicating that, at the very least, precedent is not an absolute prerequisite.
This does not mean that precedent is irrelevant. But it does mean that it falls to be evaluated in the round rather than in isolation. The Sewel convention became acknowledged as quickly as it was partly in light of political recognition of a new constitutional reality involving a degree of devolved autonomy. Meanwhile, the convention regarding Prime Minister’s Questions could be departed from as easily as it was because, clear though the precedent might have been, the change in practice did no violence to the underling norm that the Prime Minister must make himself or herself accountable to Parliament by regularly answering oral questions. More generally, it has always seemed to me that a fundamental problem with the Jennings test is that it identifies highly pertinent issues but treats them as binary matters that invite “yes” or “no” answers: “Is there a precedent?” “Do the actors feel bound?” “Is there a good constitutional reason?” The reality is messier and more complex. The questions and the answers are really ones of degree: “How strong is the precedent?” “How bound do the actors feel?” “How compelling is the justification?” And the three considerations themselves are ones that exist not in silos but in relationship with one another, the strength of the underlying norm and the extent to which it can be respected through different forms of practice serving to inform the ease and propriety with which established practices can be departed from.
On that approach, what should we make of the fact that, in 1963, Douglas-Home became Prime Minister three or so weeks before winning a by-election and taking up a seat in the Commons? Or that, the following year, Gordon Walker became Foreign Secretary in anticipation of winning a by-election (but resigned when he lost it)? Those seem to me isolated examples of deviation from the practice whereby the Prime Minister is a member of the House of Commons that, during the twentieth century, had begun to establish itself firmly as a norm – both in the sense of being what was usual and in the sense of being what was normatively expected or required. After all, the last government to be led from the House of Lords was that of Lord Salisbury, who stepped down as Prime Minister in 1902 during what can only be described, given the nature of the electoral franchise at the time, as a pre-democratic era. And when in 1923 King George V had to appoint a new Prime Minister upon the resignation of Andrew Bonar Law – in circumstances in which the premier’s illness was so grave as to preclude him from advising the King – Stanley Baldwin MP was preferred over Lord Curzon, who was regarded by many as the obvious choice. As Lord Lexden has explained, while a variety of factors might have been in play, one of them was that the King had concluded that it was “imperative that he should appoint a prime minister from the House of Commons”.
Over a century on, precedents such as Douglas-Home and Gordon Walker appear to be little more than the dying embers of a period in British constitutional history when it was acceptable for the Prime Minister to lead the government other than as a member of the House of Commons. That is arguably apparent on the basis of empirical analysis alone. But it is a conclusion that is reinforced considerably when the normative dimension is factored in. So embedded today is the principles that democratic legitimacy requires the Prime Minister to lead their government from the House of Commons, and that the Prime Minister must be accountable to that elected chamber, that any historical examples to the contrary must be viewed – with appropriate scepticism as to their contemporary relevance – through that normative lens. An equivalent point applies to examples from other countries, given the depth of commitment in the contemporary UK to the relevant norm. That does not mean that, for example, Canada’s recent practice in relation to Carney’s appointment as Prime Minister was “wrong”. But it does serve as a reminder that we should be very careful in this context before reading precedent from one jurisdiction across to another. Constitutional convention is ultimately a crystallisation of the political and constitutional morality of a given system, making it difficult, at best, to rely on practice in one system when seeking to determine what is or ought to be regarded as constitutionally legitimate in another.
It is, of course, possible to think of “edge cases”. If, for example, a Prime Minister were to resign or die during a general election campaign – when, following Parliament’s dissolution, there are no MPs – the monarch would have no choice but to appoint a new Prime Minister on an interim basis. But the fact that it is possible to think of circumstances in which it would be practically necessary (thanks to the absence of any MPs) to appoint as Prime Minister someone who was not an MP does not detract from the modern convention. When, outwith the very particular context of a general election campaign, MPs exist who can be appointed to the office of Prime Minister, it would today be unconstitutional – in the sense of being a breach of convention – for the monarch to appoint to that office someone who was not an MP. That reading of the convention accords with more than a hundred years of British constitutional history once that history is viewed, as it ought to be, through the prism of deepening normative commitment over that period to fundamental principles of democratic legitimacy and political accountability – principles that can be satisfied only by the Prime Minister’s membership of the House of Commons.
